Re: [origin tree build failure] Re: [GIT PULL] AMD64 EDAC updates, p1

[Jean Delvare <khali@linux-fr.org>]

On Tue, 15 Sep 2009 13:09:40 +0200, Ingo Molnar wrote:
> -tip testing found this build failure with the attached config:
> 
> drivers/edac/edac_mce_amd.c: In function ‘decode_mce’:
> drivers/edac/edac_mce_amd.c:408: error: ‘per_cpu__cpu_llc_id’ undeclared (first use in this function)
> drivers/edac/edac_mce_amd.c:408: error: (Each undeclared identifier is reported only once
> drivers/edac/edac_mce_amd.c:408: error: for each function it appears in.)
> 
> I did the patchlet below but this needs a cleaner fix eventually. A 
> proper node abstraction needs to be factored out and provided as a 
> facility regardless of config settings.
> 
> 	Ingo
> 
> Signed-off-by: Ingo Molnar <mingo@elte.hu>
> ---
>  drivers/edac/edac_mce_amd.c |    5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> Index: linux2/drivers/edac/edac_mce_amd.c
> ===================================================================
> --- linux2.orig/drivers/edac/edac_mce_amd.c
> +++ linux2/drivers/edac/edac_mce_amd.c
> @@ -405,8 +405,11 @@ void decode_mce(struct mce *m)
>  		regs.nbsh  = (u32)(m->status >> 32);
>  		regs.nbeal = (u32) m->addr;
>  		regs.nbeah = (u32)(m->addr >> 32);
> +#ifdef CONFIG_X86_HT
>  		node       = per_cpu(cpu_llc_id, m->extcpu);
> -
> +#else
> +		node	   = 0;
> +#endif
>  		amd_decode_nb_mce(node, &regs, 1);
>  		break;
>  

FWIW, this patch doesn't fix the problem for me. Amerigo Wang's does:
http://lkml.org/lkml/2009/9/15/64

-- 
Jean Delvare